Early prediction of ADHD symptoms from perinatal characteristics: A machine learning study

This study used machine learning on the Born in Bradford cohort to predict ADHD symptoms in children from perinatal factors including maternal health, smoking, and socioeconomic status, finding infant male sex and maternal pregnancy smoking as the strongest predictors. The model explained 6.97% of variance in ADHD symptoms, demonstrating proof-of-concept for early risk identification while acknowledging the need for additional data to improve accuracy.